When you hear the words “automated machine learning”, what comes to your mind first? For me, it’s usually H2O.ai’s Driverless AI, or Google’s Cloud AutoML. Microsoft is probably a bit down in that list (Azure, anyone?).
But that list might be about to see some changes. Microsoft has released an open-source automated machine learning toolkit on GitHub that helps a user perform neural architecture search and hyperparameter tuning. Microsoft is calling the toolkit ‘Neural Network Intelligence (NNI)’.
According to Microsoft, “the tool dispatches and runs trial jobs that generated are by tuning algorithms to search for the best neural architecture and/or hyper-parameters at different environments (e.g. local, remote servers and cloud)”. The below diagram illustrates this point well:
So who does this NNI toolkit target? And why should you consider using it (or at least giving it a go)? Use the below checklist to find out:
You can install NNI through pip by using the below command:
pip3 install -v --user git+https://github.com/Microsoft/[email protected] source ~/.bashrc
Note that you’ll need to have Python version 3.5 or greater to use this toolkit.
I’ll be honest – when I took my first steps into the dreamy world of data science, I hadn’t imagined autoML picking up so quickly. I used to hear my seniors talk about how all of that was at least 5-7 years away (back in 2016). Just goes to show how quickly technology has advanced, it’s even taken people in-the-know by surprise.
And we’re already at the stage of seeing autoML going open-source (or at least parts of it)! I covered Auto-Keras last month and that was quite a big deal in the ML community. I’m sure Microsoft’s NNI will help speed up the automated designing of models to quite an extent as well. It’s definitely worth exploring and since it’s free, there are no excuses for not doing so!
Lorem ipsum dolor sit amet, consectetur adipiscing elit,